Support all Gremlin message serializer formats for Neptune

ChangeLog.md

@@ -4,11 +4,12 @@ Starting with v1.31.6, this file will contain a record of major features and upd
 
 ## Upcoming
 
+- Updated Gremlin config `message_serializer` to accept all TinkerPop serializers ([Link to PR](https://github.com/aws/graph-notebook/pull/685))
+- Added `%get_import_task` line magic ([Link to PR](https://github.com/aws/graph-notebook/pull/668))
+- Added `--export-to` JSON file option to `%%graph_notebook_config` ([Link to PR](https://github.com/aws/graph-notebook/pull/684))
 - Deprecated Python 3.8 support ([Link to PR](https://github.com/aws/graph-notebook/pull/683))
 - Upgraded Neo4j Bolt driver to v5.x ([Link to PR](https://github.com/aws/graph-notebook/pull/682))
 - Upgraded nest_asyncio to 1.6.0 ([Link to PR](https://github.com/aws/graph-notebook/pull/698))
-- Added `%get_import_task` line magic ([Link to PR](https://github.com/aws/graph-notebook/pull/668))
-- Added `--export-to` JSON file option to `%%graph_notebook_config` ([Link to PR](https://github.com/aws/graph-notebook/pull/684))
 - Improved iPython config directory retrieval logic ([Link to PR](https://github.com/aws/graph-notebook/pull/687))
 - Fixed `%db_reset` output for token modes ([Link to PR](https://github.com/aws/graph-notebook/pull/691))
 - Fixed `%%gremlin profile` serialization issue on Neptune DB v1.2 and older ([Link to PR](https://github.com/aws/graph-notebook/pull/694))

src/graph_notebook/configuration/generate_config.py

@@ -14,7 +14,7 @@
                                            HTTP_PROTOCOL_FORMATS, WS_PROTOCOL_FORMATS,
                                            DEFAULT_NEO4J_USERNAME, DEFAULT_NEO4J_PASSWORD, DEFAULT_NEO4J_DATABASE,
                                            NEPTUNE_CONFIG_HOST_IDENTIFIERS, is_allowed_neptune_host, false_str_variants,
-                                           GRAPHSONV3_VARIANTS, GRAPHSONV2_VARIANTS, GRAPHBINARYV1_VARIANTS,
+                                           GRAPHBINARYV1, GREMLIN_SERIALIZERS_HTTP,
                                            NEPTUNE_DB_SERVICE_NAME, NEPTUNE_ANALYTICS_SERVICE_NAME,
                                            normalize_service_name)
 
@@ -73,17 +73,29 @@ def __init__(self, traversal_source: str = '', username: str = '', password: str
 
         serializer_lower = message_serializer.lower()
         # TODO: Update with untyped serializers once supported in GremlinPython
+        # Accept TinkerPop serializer class name
+        # https://github.com/apache/tinkerpop/blob/fd040c94a66516e473811fe29eaeaf4081cf104c/docs/src/reference/gremlin-applications.asciidoc#graphson
+        # https://github.com/apache/tinkerpop/blob/fd040c94a66516e473811fe29eaeaf4081cf104c/docs/src/reference/gremlin-applications.asciidoc#graphbinary
         if serializer_lower == '':
             message_serializer = DEFAULT_GREMLIN_SERIALIZER
-        elif serializer_lower in GRAPHSONV3_VARIANTS:
-            message_serializer = 'graphsonv3'
-        elif serializer_lower in GRAPHSONV2_VARIANTS:
-            message_serializer = 'graphsonv2'
-        elif serializer_lower in GRAPHBINARYV1_VARIANTS:
-            message_serializer = 'graphbinaryv1'
+        elif 'graphson' in serializer_lower:
+            message_serializer = 'GraphSON'
+            if 'untyped' in serializer_lower:
+                message_serializer += 'Untyped'
+            if 'v1' in serializer_lower:
+                if 'untyped' in serializer_lower:
+                    message_serializer += 'MessageSerializerV1'
+                else:
+                    message_serializer += 'MessageSerializerGremlinV1'
+            elif 'v2' in serializer_lower:
+                message_serializer += 'MessageSerializerV2'
+            else:
+                message_serializer += 'MessageSerializerV3'
+        elif 'graphbinary' in serializer_lower:
+            message_serializer = GRAPHBINARYV1
         else:
             print(f'Invalid Gremlin serializer specified, defaulting to graphsonv3. '
-                  f'Valid serializers: [graphsonv3, graphsonv2, graphbinaryv1].')
+                  f'Valid serializers: {GREMLIN_SERIALIZERS_HTTP}.')
             message_serializer = DEFAULT_GREMLIN_SERIALIZER
 
         self.traversal_source = traversal_source
@@ -93,16 +105,21 @@ def __init__(self, traversal_source: str = '', username: str = '', password: str
 
         if include_protocol:
             protocol_lower = connection_protocol.lower()
-            if protocol_lower == '':
-                connection_protocol = DEFAULT_GREMLIN_PROTOCOL
-            elif protocol_lower in HTTP_PROTOCOL_FORMATS:
+            if message_serializer in GREMLIN_SERIALIZERS_HTTP:
                 connection_protocol = DEFAULT_HTTP_PROTOCOL
-            elif protocol_lower in WS_PROTOCOL_FORMATS:
-                connection_protocol = DEFAULT_WS_PROTOCOL
+                if protocol_lower != '' and protocol_lower not in HTTP_PROTOCOL_FORMATS:
+                    print(f"Enforcing HTTP protocol usage for serializer: {message_serializer}.")
             else:
-                print(f"Invalid connection protocol specified, defaulting to {DEFAULT_GREMLIN_PROTOCOL}. "
-                      f"Valid protocols: [websockets, http].")
-                connection_protocol = DEFAULT_GREMLIN_PROTOCOL
+                if protocol_lower == '':
+                    connection_protocol = DEFAULT_GREMLIN_PROTOCOL
+                elif protocol_lower in HTTP_PROTOCOL_FORMATS:
+                    connection_protocol = DEFAULT_HTTP_PROTOCOL
+                elif protocol_lower in WS_PROTOCOL_FORMATS:
+                    connection_protocol = DEFAULT_WS_PROTOCOL
+                else:
+                    print(f"Invalid connection protocol specified, defaulting to {DEFAULT_GREMLIN_PROTOCOL}. "
+                          f"Valid protocols: [websockets, http].")
+                    connection_protocol = DEFAULT_GREMLIN_PROTOCOL
             self.connection_protocol = connection_protocol
 
     def to_dict(self):

src/graph_notebook/magics/graph_magic.py

@@ -52,10 +52,10 @@
     NEPTUNE_CONFIG_HOST_IDENTIFIERS, is_allowed_neptune_host, \
     STATISTICS_LANGUAGE_INPUTS, STATISTICS_LANGUAGE_INPUTS_SPARQL, STATISTICS_MODES, SUMMARY_MODES, \
     SPARQL_EXPLAIN_MODES, OPENCYPHER_EXPLAIN_MODES, GREMLIN_EXPLAIN_MODES, \
-    OPENCYPHER_PLAN_CACHE_MODES, OPENCYPHER_DEFAULT_TIMEOUT, OPENCYPHER_STATUS_STATE_MODES,
+    OPENCYPHER_PLAN_CACHE_MODES, OPENCYPHER_DEFAULT_TIMEOUT, OPENCYPHER_STATUS_STATE_MODES, \
     normalize_service_name, NEPTUNE_DB_SERVICE_NAME, NEPTUNE_ANALYTICS_SERVICE_NAME, GRAPH_PG_INFO_METRICS, \
-    DEFAULT_GREMLIN_PROTOCOL, GREMLIN_PROTOCOL_FORMATS, DEFAULT_HTTP_PROTOCOL, normalize_protocol_name,
-    generate_snapshot_name)
+    DEFAULT_GREMLIN_PROTOCOL, GREMLIN_PROTOCOL_FORMATS, DEFAULT_HTTP_PROTOCOL, DEFAULT_WS_PROTOCOL, \
+    GREMLIN_SERIALIZERS_WS, GREMLIN_SERIALIZERS_CLASS_TO_MIME_MAP, normalize_protocol_name, generate_snapshot_name)
 from graph_notebook.network import SPARQLNetwork
 from graph_notebook.network.gremlin.GremlinNetwork import parse_pattern_list_str, GremlinNetwork
 from graph_notebook.visualization.rows_and_columns import sparql_get_rows_and_columns, opencypher_get_rows_and_columns
@@ -1249,14 +1249,23 @@ def gremlin(self, line, cell, local_ns: dict = None):
             using_http = False
             query_start = time.time() * 1000  # time.time() returns time in seconds w/high precision; x1000 to get in ms
             if self.client.is_neptune_domain():
-                connection_protocol = normalize_protocol_name(args.connection_protocol) \
-                    if args.connection_protocol != '' \
-                    else self.graph_notebook_config.gremlin.connection_protocol
+                if args.connection_protocol != '':
+                    connection_protocol = normalize_protocol_name(args.connection_protocol)
+                    if connection_protocol == DEFAULT_WS_PROTOCOL and \
+                            self.graph_notebook_config.gremlin.message_serializer not in GREMLIN_SERIALIZERS_WS:
+                        print("Unsupported serializer for GremlinPython client, "
+                              "compatible serializers are: {GREMLIN_SERIALIZERS_WS}")
+                        print("Defaulting to HTTP protocol.")
+                        connection_protocol = DEFAULT_HTTP_PROTOCOL
+                else:
+                    connection_protocol = self.graph_notebook_config.gremlin.connection_protocol
                 try:
                     if connection_protocol == DEFAULT_HTTP_PROTOCOL:
                         using_http = True
+                        message_serializer = self.graph_notebook_config.gremlin.message_serializer
+                        message_serializer_mime = GREMLIN_SERIALIZERS_CLASS_TO_MIME_MAP[message_serializer]
                         query_res_http = self.client.gremlin_http_query(cell, headers={
-                            'Accept': 'application/vnd.gremlin-v1.0+json;types=false'})
+                            'Accept': message_serializer_mime})
                         query_res_http.raise_for_status()
                         try:
                             query_res_http_json = query_res_http.json()

src/graph_notebook/neptune/client.py

@@ -32,7 +32,6 @@
 # client >= 3.5.0 as the HashableDict is now part of that client driver.
 # import graph_notebook.neptune.gremlin.graphsonV3d0_MapType_objectify_patch  # noqa F401
 
-DEFAULT_GREMLIN_SERIALIZER = 'graphsonv3'
 DEFAULT_GREMLIN_TRAVERSAL_SOURCE = 'g'
 DEFAULT_SPARQL_CONTENT_TYPE = 'application/x-www-form-urlencoded'
 DEFAULT_PORT = 8182
@@ -119,16 +118,35 @@
 
 false_str_variants = [False, 'False', 'false', 'FALSE']
 
-GRAPHSONV3_VARIANTS = ['graphsonv3', 'graphsonv3d0', 'graphsonserializersv3d0', 'graphsonmessageserializerv3']
-GRAPHSONV2_VARIANTS = ['graphsonv2', 'graphsonv2d0', 'graphsonserializersv2d0', 'graphsonmessageserializerv2']
-GRAPHBINARYV1_VARIANTS = ['graphbinaryv1', 'graphbinary', 'graphbinaryserializersv1', 'graphbinarymessageserializerv1']
+GRAPHSONV1 = 'GraphSONMessageSerializerGremlinV1'
+GRAPHSONV2 = 'GraphSONMessageSerializerV2'
+GRAPHSONV3 = 'GraphSONMessageSerializerV3'
+GRAPHSONV1_UNTYPED = 'GraphSONUntypedMessageSerializerV1'
+GRAPHSONV2_UNTYPED = 'GraphSONUntypedMessageSerializerV2'
+GRAPHSONV3_UNTYPED = 'GraphSONUntypedMessageSerializerV3'
+GRAPHBINARYV1 = 'GraphBinaryMessageSerializerV1'
+
+GREMLIN_SERIALIZERS_CLASS_TO_MIME_MAP = {
+    GRAPHSONV1: 'application/vnd.gremlin-v1.0+json',
+    GRAPHSONV2: 'application/vnd.gremlin-v2.0+json',
+    GRAPHSONV3: 'application/vnd.gremlin-v3.0+json',
+    GRAPHSONV1_UNTYPED: 'application/vnd.gremlin-v1.0+json;types=false',
+    GRAPHSONV2_UNTYPED: 'application/vnd.gremlin-v2.0+json;types=false',
+    GRAPHSONV3_UNTYPED: 'application/vnd.gremlin-v3.0+json;types=false',
+    GRAPHBINARYV1: 'application/vnd.graphbinary-v1.0'
+}
+
+GREMLIN_SERIALIZERS_WS = [GRAPHSONV2, GRAPHSONV3, GRAPHBINARYV1]
+GREMLIN_SERIALIZERS_HTTP = [GRAPHSONV1, GRAPHSONV1_UNTYPED, GRAPHSONV2_UNTYPED, GRAPHSONV3_UNTYPED]
+GREMLIN_SERIALIZERS_ALL = GREMLIN_SERIALIZERS_WS + GREMLIN_SERIALIZERS_HTTP
+DEFAULT_GREMLIN_SERIALIZER = GRAPHSONV1_UNTYPED
 
 DEFAULT_WS_PROTOCOL = "websockets"
 DEFAULT_HTTP_PROTOCOL = "http"
 WS_PROTOCOL_FORMATS = ["ws", "websocket", DEFAULT_WS_PROTOCOL]
 HTTP_PROTOCOL_FORMATS = ["https", "rest", DEFAULT_HTTP_PROTOCOL]
 GREMLIN_PROTOCOL_FORMATS = WS_PROTOCOL_FORMATS + HTTP_PROTOCOL_FORMATS
-DEFAULT_GREMLIN_PROTOCOL = DEFAULT_WS_PROTOCOL
+DEFAULT_GREMLIN_PROTOCOL = DEFAULT_HTTP_PROTOCOL
 
 STATISTICS_MODES = ["", "status", "disableAutoCompute", "enableAutoCompute", "refresh", "delete"]
 SUMMARY_MODES = ["", "basic", "detailed"]
@@ -153,16 +171,22 @@ def is_allowed_neptune_host(hostname: str, host_allowlist: list):
     return False
 
 
-def get_gremlin_serializer(serializer_str: str):
-    serializer_lower = serializer_str.lower()
-    if serializer_lower == 'graphbinaryv1':
+def get_gremlin_serializer_driver_class(serializer_str: str):
+    if serializer_str == GRAPHBINARYV1:
         return serializer.GraphBinarySerializersV1()
-    elif serializer_lower == 'graphsonv2':
+    elif serializer_str == GRAPHSONV2:
         return serializer.GraphSONSerializersV2d0()
     else:
         return serializer.GraphSONSerializersV3d0()
 
 
+def get_gremlin_serializer_mime(serializer_str: str):
+    if serializer_str in GREMLIN_SERIALIZERS_CLASS_TO_MIME_MAP.keys():
+        return GREMLIN_SERIALIZERS_CLASS_TO_MIME_MAP[serializer_str]
+    else:
+        return GREMLIN_SERIALIZERS_CLASS_TO_MIME_MAP[GRAPHSONV1_UNTYPED]
+
+
 def normalize_protocol_name(protocol: str):
     if protocol in WS_PROTOCOL_FORMATS:
         return DEFAULT_WS_PROTOCOL
@@ -223,7 +247,7 @@ def __init__(self, host: str, port: int = DEFAULT_PORT,
         self.gremlin_traversal_source = gremlin_traversal_source
         self.gremlin_username = gremlin_username
         self.gremlin_password = gremlin_password
-        self.gremlin_serializer = get_gremlin_serializer(gremlin_serializer)
+        self.gremlin_serializer = gremlin_serializer
         self.neo4j_username = neo4j_username
         self.neo4j_password = neo4j_password
         self.neo4j_auth = neo4j_auth
@@ -373,9 +397,10 @@ def get_gremlin_connection(self, transport_kwargs) -> client.Client:
             request = self._prepare_request('GET', ws_url)
 
         traversal_source = 'g' if self.is_neptune_domain() else self.gremlin_traversal_source
+        message_serializer = get_gremlin_serializer_driver_class(self.gremlin_serializer)
         return client.Client(ws_url, traversal_source, transport_factory=transport_factory_args,
                              username=self.gremlin_username, password=self.gremlin_password,
-                             message_serializer=self.gremlin_serializer,
+                             message_serializer=message_serializer,
                              headers=dict(request.headers), **transport_kwargs)
 
     def gremlin_query(self, query, transport_args=None, bindings=None):

test/unit/configuration/test_configuration.py

@@ -51,7 +51,7 @@ def test_generate_default_config(self):
         self.assertEqual('', config.gremlin.username)
         self.assertEqual('', config.gremlin.password)
         self.assertEqual(DEFAULT_GREMLIN_PROTOCOL, config.gremlin.connection_protocol)
-        self.assertEqual('graphsonv3', config.gremlin.message_serializer)
+        self.assertEqual('GraphSONUntypedMessageSerializerV1', config.gremlin.message_serializer)
         self.assertEqual('neo4j', config.neo4j.username)
         self.assertEqual('password', config.neo4j.password)
         self.assertEqual(True, config.neo4j.auth)
@@ -170,7 +170,7 @@ def test_get_configuration_generic_required_input(self):
                 'traversal_source': 'g',
                 'username': '',
                 'password': '',
-                'message_serializer': 'graphsonv3'
+                'message_serializer': 'GraphSONUntypedMessageSerializerV1'
             },
             'neo4j': {
                 'username': 'neo4j',
@@ -197,7 +197,7 @@ def test_get_configuration_generic_all_input(self):
                 'traversal_source': 'a',
                 'username': 'user',
                 'password': 'pass',
-                'message_serializer': 'graphbinaryv1'
+                'message_serializer': 'GraphBinaryMessageSerializerV1'
             },
             'neo4j': {
                 'username': 'neo_user',
@@ -267,8 +267,8 @@ def test_get_configuration_neptune_required_input(self):
                 'traversal_source': 'g',
                 'username': '',
                 'password': '',
-                'message_serializer': 'graphsonv3',
-                'connection_protocol': 'websockets'
+                'message_serializer': 'GraphSONUntypedMessageSerializerV1',
+                'connection_protocol': 'http'
             },
             'neo4j': {
                 'username': 'neo4j',
@@ -300,7 +300,7 @@ def test_get_configuration_neptune_all_input(self):
                 'traversal_source': 'a',
                 'username': 'a_user',
                 'password': 'a_pass',
-                'message_serializer': 'graphbinaryv1',
+                'message_serializer': 'GraphSONUntypedMessageSerializerV3',
                 'connection_protocol': 'http'
             },
             'neo4j': {
@@ -328,7 +328,7 @@ def test_get_configuration_neptune_all_input(self):
                 'traversal_source': 'g',
                 'username': '',
                 'password': '',
-                'message_serializer': 'graphbinaryv1',
+                'message_serializer': 'GraphSONUntypedMessageSerializerV3',
                 'connection_protocol': 'http'
             },
             'neo4j': {
@@ -472,7 +472,7 @@ def test_configuration_gremlinsection_generic_default(self):
         self.assertEqual(config.gremlin.traversal_source, 'g')
         self.assertEqual(config.gremlin.username, '')
         self.assertEqual(config.gremlin.password, '')
-        self.assertEqual(config.gremlin.message_serializer, 'graphsonv3')
+        self.assertEqual(config.gremlin.message_serializer, 'GraphSONUntypedMessageSerializerV1')
         self.assertFalse(hasattr(config.gremlin, "connection_protocol"))
 
     def test_configuration_gremlinsection_generic_override(self):
@@ -486,15 +486,15 @@ def test_configuration_gremlinsection_generic_override(self):
         self.assertEqual(config.gremlin.traversal_source, 't')
         self.assertEqual(config.gremlin.username, 'foo')
         self.assertEqual(config.gremlin.password, 'bar')
-        self.assertEqual(config.gremlin.message_serializer, 'graphbinaryv1')
+        self.assertEqual(config.gremlin.message_serializer, 'GraphBinaryMessageSerializerV1')
         self.assertFalse(hasattr(config.gremlin, "connection_protocol"))
 
     def test_configuration_gremlinsection_neptune_default(self):
         config = Configuration(self.neptune_host_reg, self.port)
         self.assertEqual(config.gremlin.traversal_source, 'g')
         self.assertEqual(config.gremlin.username, '')
         self.assertEqual(config.gremlin.password, '')
-        self.assertEqual(config.gremlin.message_serializer, 'graphsonv3')
+        self.assertEqual(config.gremlin.message_serializer, 'GraphSONUntypedMessageSerializerV1')
         self.assertEqual(config.gremlin.connection_protocol, DEFAULT_GREMLIN_PROTOCOL)
 
     def test_configuration_gremlinsection_neptune_override(self):
@@ -510,7 +510,7 @@ def test_configuration_gremlinsection_neptune_override(self):
         self.assertEqual(config.gremlin.traversal_source, 'g')
         self.assertEqual(config.gremlin.username, '')
         self.assertEqual(config.gremlin.password, '')
-        self.assertEqual(config.gremlin.message_serializer, 'graphbinaryv1')
+        self.assertEqual(config.gremlin.message_serializer, 'GraphBinaryMessageSerializerV1')
         self.assertEqual(config.gremlin.connection_protocol, DEFAULT_HTTP_PROTOCOL)
 
     def test_configuration_gremlinsection_protocol_neptune_default_with_proxy(self):